I also found Ming Thein's article intriguing. He has found some of same rough edges around the HR mode that I have run into. Making setting changes while in HR mode has unexpected side effects. I tried put HR parameters into a Myset and putting the Myset on the HDR button to get a one-button path into and back out of HR, but it doesn't always work. A better procedure is to just go into the menu, turn HR on and select a time delay. Turn it off when you are done. I expect that firmware changes on the M5.2 will fix some of this (and add new surprises).
But Ming complains that the raw HD files, viewed at the pixel level, are soft. Of course they are, since each of the new estimated 64 MPx "virtual pixels" is an estimate based on information gathered from an area half again as large as the original pixel. So sharpening will be required to get acutance, and it may be that it only comes at the expense of nice tone modulation. Also, a subtle trick that Sandy McGuffog noticed with his AccuRaw is that Olympus writes a raw file that pretends to have a Bayer filter array on the 64MPx spatial scale, so Moire, which could have been cancelled out, reappears on the smaller scale. The benefit of this trick is that existing software can process the 64MPx files without a murmur.
scott